عنوان مقاله [English]
نویسندگان [English]چکیده [English]
Precipitation is a vital component of the global water and energy cycle with large variations in space and time. The observational datasets that are based on meteorological stations data usually serve as the main sources of precipitation data. However, because of uneven distribution in space, such datasets may not be directly applicable to some problems. Furthermore, there are gaps in the data as there may be times for which the precipitation has not been recorded by some metrological stations for various technical reasons. In the last decades, several gridded precipitation databases have been developed by researchers or institutes. The main aim of creating these databases is to serve user requirements and solve the problems mentioned above. The even distribution in space of the gridded precipitation data and their availability are two very important factors. These databases are critical for many studies including climate change and numerical weather prediction (NWP) applications, management of water resources, agriculture, and disaster management.
The Global Precipitation Climatology Centre (GPCC) has been established in the year 1989 at the request of the World Meteorological Organization (WMO). It is constructed by the Deutscher Wetterdienst (DWD, National Meteorological Service of Germany) as a German contribution to the World Climate Research Programme (WCRP). The precipitation data of GPCC are freely available via the website http://gpcc.dwd.de at 2.5º × 2.5º, 1º×1º, and 0.5 º × 0.5 º resolutions.
The aim of this research is to evaluate the accuracy of GPCC database over Iran by comparing it with two national databases, the Asfezari and that of the synoptic stations called Stations hereafter. The monthly precipitation data from the three databases including GPCC, Asfezari and Stations have been used from January 1962 to the end of December 2010. To evaluate the accuracy of the estimated GPCC precipitation data, first the spatial resolutions of the three databases have been synchronized by the nearest neighbor algorithm. The high-resolution database is converted to the low-resolution database in order to select spatial pixels and carry out the comparisons. Seven accuracy evaluation indices have been used.
The results indicate a high temporal correlation between the estimated precipitation of GPCC and the observed precipitation by the Asefazri and the Stations databases. The results of applying accuracy evaluation indices to the precipitation time series show that in addition to high temporal correlation, quantitatively the estimated precipitations are also very similar to the observed precipitations. Although in some regions the estimated precipitation values are contaminated with bias, but overall the estimated precipitation error is low compared to the total precipitation received. In a spatial viewpoint, the highest accuracy is observed over the western parts of Zagros mountain range and the northeast of the country. Over these regions, the index of agreement and the coefficient of determination are close to unity. The highest relative root mean square (rms) is observed over the dry interior regions and the Lut desert. The relative rms is low in the regions with high precipitation when compared with the rest of the regions. In a temporal viewpoint, the highest correlation between the precipitation time series is observed in rainy months. Based on the results from the Nash–Sutcliffe efficiency index, over most regions of the country, using the estimated precipitations by GPCC is preferable to applying the mean precipitation amounts. The results of this study confirmed the finding of other researches about the accuracy of the estimated precipitation by GPCC database.